Vendors tackle the problems of where to store, when to access, and how to archive data.
By Mike May
Oct. 8, 2008 | As scientists grapple with the deluge of data in virtually every facet of drug R&D, data storage companies are themselves racing to offer the best and greenest solutions (see, “The DNA Data Deluge,” Bio-IT World, April 2008). Beyond the difficulty of simply storing that data somewhere, scientists need ways to work with it. A spate of offerings from leading data storage vendors addresses those issues and reveals that many of today’s decisions depend largely on timing.
The best place to store data depends on when it will be needed as well as its age. It usually makes sense to put fresh data on fast-access storage, like a disk, and older data relegated to tape. For example, Deepak Thakkar, director of industry marketing at Silicon Graphics Inc. (SGI) in Sunnyvale, California, says that with SGI Data Migration Facility software, “we can orchestrate incoming data into different storage channels.” With this software, a user defines rules that automatically organize stored data.
Tape storage is a must in some circumstances. Bruce Hillsberg, IBMs’ director of storage system research in San Jose, California, says, “If you are storing billions of files and petabytes of data—and companies do that—then you need an effective and green way of doing that. Tape is a great choice.” SGI also supports the greening trend in today’s data storage. For example, when SGI’s data migration facility puts information on tape, that storage requires zero power, unless the data are being accessed. So while today’s scientists collect more information, they can (sometimes) do it with less energy.
Many of the biggest improvements in data-storage capabilities come from hardware-software combinations. In August, EMC (Hopkinton, Massachusetts) introduced its CLARiiON CX4 series, which feeds off several advanced technologies. For one thing, the CX4 series uses flash drives, which provide nonspinning, solid-state storage. According to Ruya Atac-Barrett, EMC’s director of CLARiiON marketing, “These drives provide 30 times more processing power than spinning drives and response times are 10 times faster.” She adds, “Flash drives are 98% energy efficient.” One CLARiiON system can also store nearly a petabyte of data. Moreover, RSA enVision security software comes built-in with CLARiiON CX4 systems. This software package is produced by EMC’s security division, RSA, and it provides a range of features, including reporting unusual usage and summaries of inbound and outbound traffic.
Also in August, EMC released Celerra NX4, which is aimed at medium-size health care organizations. It can be used in a variety of configurations: network attached storage (NAS), Internet smart computer system interface (iSCSI), and fiber-channel storage area networks (SAN). The NX4 also provides 60 terabytes of storage.
In many life science research applications, changing technology requires data storage that can grow. Consequently, BlueArc (San Jose, California) adds Dynamic Write Balancing to its storage solutions. “When you add storage,” says James Reaney, BlueArc’s director of research markets, “this automatically rebalances the data across old and new storage. Otherwise, you would have to copy all of that data, reformat it, and then copy it back into storage.”
Isilon Systems in Seattle, Washington, also kept growth in mind with its IQ X Series. “You can start with a simple system, and then add nodes as you need them,” says Jay Wampold, Isilon’s senior director of marketing and communications. “When you add a node to the cluster, the system automatically balances the data.” In fact, Isilon says that its storage systems are so easy to use that it takes less work to manage data. Chris Blessington, Isilon’s director of marketing communications, says that one customer “had been using six full-time IT people just to manage storage. After switching to an Isilon system, they needed zero full-time employees to manage storage.” (see, “Isilon Insights,” Bio-IT World, April 2008).
In September, Isilon introduced an even more-advanced storage system, its OneFS (one file system) 5.0. “This delivers 20 gigabytes of aggregate throughput,” says Wampold, “and it can scale up to 2.3 petabytes of storage in a single-file system.
Archiving for the Ages
There are certain cases where biotechnology and pharmaceutical companies need to keep some data for years, possibly even decades. Much of that data might include repeats, such as reports that use some of the same data. In addition, those data must be kept safe, even though they won’t be used often. IBM provides a few new approaches to those problems.
For replicated data in stored files, researchers can turn to deduplication, which IBM’s Hillsberg describes as “the ability to take data that you will store, look for repeating patterns, and then factor out those patterns.” To move into deduplication faster, IBM acquired Diligent Technologies in Framingham, Massachusetts. “Diligent has the leading deduplication product, and it can reduce the space that information needs by a factor of 25, depending on the data,” says Hillsberg.
In July, IBM introduced its System Storage TS1130, which holds up to a terabyte of data. And for a tape drive, this one is fast—providing an input-output bandwidth of 160 megabytes per second. The tapes used for this system are smaller than two packs of playing cards. Moreover, data put on one of these tapes should last a decade, maybe longer. “It depends how you care for the tape,” says Hillsberg. In addition, the TS1130 even includes data encryption. “If you lose a tape,” Hillsberg says, “there’s no way anybody could get data off it.”
Despite the speed of the TS1130, it’s not like disk storage in terms of access. “For data that you need in subseconds, store it on a disk,” Hillsberg says. But how do you decide which data to put where? If desired, IBM’s DR550 storage system will decide for you. “It comes with embedded disks, and it can use disk or tape storage that you already have in your storage network,” Hillsberg says. Likewise, BlueArc’s Data Migrator runs on a set of user-defined rules to decide where to store data. “This is very useful to get the data to where all of the researchers are,” says Reaney.
Interacting with Data
“We tend to think of storage as just storage,” says Thakkar, but he sees the future of storage as a collaborative entity. Imagine data stored in some virtual environment where many people could access the data simultaneously. “People could collaborate with each other’s data in real time, comment on it, and then merge together different data sets,” he says. “If everyone just works in their own silos then drug-research programs are only as good as one person’s thought processes. If we could collaborate, that’s good for all of us.”
Another task ahead involves so-called logical preservation, which means being able to open a file that is decades old and still having software to read and manipulate the data. “IBM is working on this in our research division,” says Hillsberg. “It requires figuring out the best way to do this without losing data because of changes in applications and formats.”
So time plays a fundamental role in a range of data-storage characteristics. Time impacts the best format for storage, and as time passes, the data must remain safe and readable. Yet, staying on top of the rushing ticks of data collected keeps storage scientists very busy.
This article appeared in Bio-IT World Magazine.
Subscriptions are free for qualifying individuals. Apply Today.